PUBLISHED ON 21 MAY 2026
— attributed to inventor Charles Kettering
— US Supreme Court Justice Oliver Wendell Holmes Jr.
Countries are racing to ban social media for children, but there’s a basic uncertainty at the core of these proposals: what does “social media” actually mean? And what is it that we’re actually banning?
While everyone has some idea of social media and its risks, there's no accepted common definition in the debate, and using “social media” as a placeholder is problematic. The gap in definitions is one reason why policymakers can indulge in ban-solutionism: banning something that is vaguely understood is easier than trying to precisely define a complex issue. But without agreeing on what it is we’re talking about, there can only be vibe politics, not a sustainable solution.
To understand this, Australia and the United Kingdom present interesting case studies. Australia was the first country to implement a social media ban for children at the end of 2025. However, it faces criticism over poor definitions which are unable to keep up with the ever-evolving digital landscape, and the ban does not seem to have worked well in practice so far.
Meanwhile the UK looks close to following suit, with a new law granting the Government the power to prevent or restrict children’s access to “specific internet services”. Both countries are being closely watched by others, with the risk of copycat proposals looming.
Australian law now says that “age-restricted social media platforms must take reasonable steps to prevent [under-16s] having accounts”. In effect, this law bans under-16s from having accounts on certain platforms.
“But which platforms?”, we hear you ask. Let’s look at the law’s definition and try to understand how it fits with our general understanding of what social media is.
The definition covers electronic services that meet three conditions: 1) that their sole purpose, or a significant purpose, is to enable online social interaction between two or more end-users; 2) that they allow end‑users to link to, or interact with, some or all of the other end‑users; and 3) that they allow end‑users to post material on the service.
The law gives the Minister for Communications significant power to set rules over how the ban applies, including to decide what further features and characteristics of platforms result in them being age restricted. This power was exercised most recently in March 2026. The Minister can also decide whether certain electronic services are included in or excluded from the ban. There are some limits on how the Minister can use this power. For example, they must seek and have regard to the advice of the eSafety Commissioner - Australia’s regulator responsible for monitoring compliance.
As we’ll see below with the UK, too, this kind of power is controversial because it allows a minister who wasn’t directly elected by citizens to amend a law made by the elected Parliament.
Ultimately, it’s Australian courts who determine whether a specific platform is or isn’t age-restricted under the law. However, eSafety has encouraged self-assessment by platforms and has tried to provide some clarity. It has published a list of ten platforms which, in its view, are age-restricted, and ten platforms which are not.
Let’s see if we can guess which is which.
A photo and video sharing social networking service. An online game platform where users programme and play games they’ve created. A video live-streaming service for content on gaming and more. And a video sharing platform. Which is the odd one out?
The correct answer is the second one, Roblox, as it’s the only one out of the four which is not age-restricted. Don’t worry if you didn’t get it right. Many don’t, and for good reason.
According to eSafety, the following platforms are banned for under-16s: Facebook, Instagram, Kick, Reddit, Snapchat, Threads, TikTok, Twitch, X and YouTube.
The following are not: Discord, GitHub, Google Classroom, LEGO Play, Messenger, Pinterest, Roblox, Steam and Steam Chat, WhatsApp and YouTube Kids.
The no-ban list can certainly seem quite arbitrary, as it includes platforms which many of us would intuitively think of as “social media”. Looking at the law itself, Roblox, for instance, one of its significant purposes is to enable online social interaction, which allows users to link with others, and to post material.
This isn’t to say that Roblox should also be banned, but there’s a puzzling inconsistency. The rules made by the Minister exclude platforms whose “sole or primary purpose” is enabling messaging or online gaming. But for some experts, the omission of gaming platforms, “doesn’t make much sense”, as “gaming and social media are so interconnected”, for example through live stream gaming.
And here’s the catch: eSafety recognises that many platforms have multiple purposes, the way platforms are used can change over time, and new platforms are constantly popping up. For instance, if a messaging or online gaming service has “social-media style features” and its primary purpose changes due to people’s “noticeable common use” of these features, it may become age-restricted.
There’s a real possibility that this approach will turn into a game of whack-a-mole, rendering the ban inadequate as a structural solution. And if developments since the start of the ban are anything to go by, it already has.
Under-16s grasped the idea of change barely a month into the ban, with many migrating to unbanned sites such as Roblox, Discord and Steam, and a how-to-skirt-the-ban culture emerged. Fast forward to the end of April 2026, following media reports about grooming and terrorism- and violent extremist-themed gameplay on four major online gaming platforms (Roblox, Minecraft, Fortnite and Steam), eSafety issued legally enforceable transparency notices asking these platforms to explain how they are protecting children. That same month, Roblox, announced that it would be introducing new age-based accounts and expanding parental controls for under-16s globally. Importantly, the company did this without first being placed on the banned list - indeed, it probably made the move to prevent this from happening. But the situation proves that product design measures can be taken without imposing a ban.
While it’s too soon to definitively say, studies so far point to a challenging picture. Children are migrating to unbanned platforms where reports of harm are increasing (as we saw above), and eSafety’s own compliance update from March 2026 found that many under-16s still have their accounts, or can create new ones, on banned platforms. Also, there was no drop in the number of under-16s reporting harm to eSafety. In a survey conducted by eSafety in January 2026, around 70% of parents said their children still had a Facebook, Instagram, Snapchat or TikTok account, and around 50% said their children still had a YouTube account. eSafety is investigating non-compliance by these five platforms.
Molly Rose Foundation, a UK-based charity that works to prevent suicide in young people by advocating for online safety, conducted the first large-scale polling of Australians aged 12-15 on the ban. It found that 61% continued to have accounts on banned platforms and 70% said it was easy to circumvent the ban. Regarding online safety, 51% said the ban made no difference, while 14% reported feeling less safe after the ban.
Molly Rose Foundation warned that given the findings, it would be a “high stakes gamble” for the UK to follow suit.
Now the UK is embroiled in its own, quite confusing, approach.
On 29 April 2026, the UK Parliament adopted a law giving a senior government minister sweeping powers to prevent and restrict children from certain internet services. This was the culmination of a tumultuous legislative process, during which an outright social media ban for under-16s was rejected three times.
These powers are controversial because, similarly to Australia, the minister can amend or repeal legislation made by the elected Parliament.
This law comes at a particularly significant moment. The UK Government has been carrying out a national consultation on children’s online safety, which ends on 26 May. Passing this law before all of the answers to the consultation have even been received (let alone analysed in depth) raises the question of whether the whole process is a foregone conclusion. As a junior minister said revealingly: “We are consulting on the mechanism, which is the right thing to do, but we are clear that under any outcome we will impose some form of age or functionality [restriction] on children under 16”.
Under the law, the minister has the power to introduce regulations that require “providers of specified internet services” to prevent or restrict children’s access to these services or to their “specific functionalities or other features”.
“But what’s the difference between preventing and restricting access to services?” is the natural question. Preventing access seems to mean denying access to children, while restricting access can refer to limiting the amount of time or the times at which children can gain access.
It’s clear that the law envisages some measures will be taken in light of the consultation. Indeed, the minister “must exercise the power” to make the provisions they consider appropriate following the end of the consultation, and they “must have regard to the responses” to the consultation.
The minister has to move quickly. By 29 July 2026, they must present a progress statement to Parliament, then have a year to put the first regulations before Parliament. If this is not done in time, they must explain why and then have six months to present those regulations.
It’s a tight deadline for such far-reaching measures. So what does the consultation, which the minister must have regard to, reveal about the next steps?
First things first: the public consultation does not define social media. It does, however, pose a range of questions about how to regulate a wide variety of services.
Which platforms?
The consultation starts by looking at the issue in a misguided way. Instead of recognising the online space as an environment where children naturally belong and where their rights automatically apply, it seems to weigh the pros and cons of children being online in the first place.
One of the first questions asked is: “Do you think the benefits of children using social media, and being online [emphasis added], outweigh the risks, or the other way around?” This simplistic framing is not consistent with a rights-based approach. As we’ve explained before, restrictions on children’s rights online are not a matter of the risks outweighing the benefits; they are a matter of necessity and proportionality. This means that they must be effective in achieving their aim (for example, of reducing harms or risks) and they must also be the least restrictive way to do so. Respondents are not given this context.
Even though the Government acknowledges that “there is no straightforward definition of social media”, the consultation goes on to ask whether respondents support minimum age restrictions on “social media services” and, if so, at what age.
More broadly, seeming to learn from the Australian experience, the UK consultation recognises that “if minimum age restrictions only apply to a small number of the most popular user-to-user services, this could risk displacing children into other online spaces, including less regulated ones”.
So later in the questionnaire, it seeks views on what factors are important for determining which apps, sites or services should have a minimum age of access or age restrictions on specific features and functionalities (on the latter, see more below).
The Government is effectively consulting on “whether the definition should: align with the Australian definition; include additional conditions based on functionality or purpose; capture services such as some gaming sites and messaging services that might have very similar functionality and risk profiles as those sites traditionally thought of as ‘social media'; include exemptions to avoid capturing low risk and educational services”.
What design features?
In contrast to Australia, the UK consultation very clearly considers bans on particular features and functionalities of services for children.
It seeks views on whether there should be minimum age restrictions and, if so, at what age, for: live streaming, the ability to send nude images or videos, disappearing content, location sharing, and connecting or talking to strangers.
A separate section is dedicated to the excessive use of technology. It asks what the minimum ages should be for “persuasive” design features: infinite scrolling, autoplay, affirmation features such as likes and comments, alerts and push notifications, as well as content recommendation algorithms. It also asks about daily screen limits and restricting overnight access for individual apps.
It’s quite surprising that the Government entertains the idea of overnight access restrictions, as the idea was implemented before elsewhere - unsuccessfully. In 2021, South Korea introduced a ban on online gaming for children between midnight and 6 AM. Although the ban initially reduced children’s time spent online, within four years their internet use had increased. Studies found practically no effect on time spent online (internet addiction declined by only 0.7%), on academic performance, or on sleep time (which increased by only 1.5 minutes per child). South Korea ended their policy.
Beyond social media: AI chatbots
Again, more far-reaching than the Australian approach, the UK consultation also covers AI chatbots. It asks how children benefit from using them, and which features are most risky to children. There is a long selection, including the realism of interactions, the personalisation of interactions, how chatbots mimic relationships and empathy, flattering language used, features to encourage more questions, the problem of hallucinations, and the ability to engage in or generate mature (sexual/romantic) content.
The consultation asks which functionalities of AI chatbots should be age-restricted, and whether all AI chatbots should be banned for those under a certain age.
Australia and the UK have both adopted a restrictionist mentality without first being able to provide a clear or satisfactory definition of social media. While Australia has a narrower approach - banning particular services that fulfil certain criteria - the UK seems poised to take much more complex and far-reaching measures. This could include restricting certain functionalities and features with different age limits, which would create a complicated online landscape.
Both countries are kicking the can down the road by giving vast powers to ministers rather than parliament to control the boundaries of as-yet-undefined social media. Compared to a parliamentary process, this affords less democratic accountability, debate and scrutiny, which increases the potential for overreach and jeopardy to children’s fundamental rights.
It’s significant that neither country has had a comprehensive discussion on the impact of social media regulation of the full range of children’s rights.
Data from Australia so far seems to suggest that the ban has had mixed results at best. Whether children there migrate to unbanned platforms, keep their existing accounts or create new ones on banned platforms, or circumvent the ban, for example by using VPNs, many do not seem to trust the restrictionist logic.
Aware of this, the UK asks questions around effective compliance and enforcement of restrictions, including by asking whether children’s access to VPNs should be restricted as well. This will be the focus of our next debate.
This piece is co-published with Defend Digital Me.
Sign up to our newsletter to stay updated and follow us on LinkedIn.
Related content: Issues - Digital rights and Exploring children’s rights in the digital environment
